|
Abstract:
|
Survey mode effects refer to systematic variations in the data obtained using different modes of data collection. Mode effects have been widely reported in the literature, but few mode effects studies have focused on the military population, whose demographics differ from the general population. Furthermore, military surveys typically involve questions that are unique to the military and military families. As multimode surveys are becoming the norm, it is important to understand if and how survey mode may influence sample estimates. Two recent Army surveys of military populations adopted a mixed mode approach involving mail and Web-based data collections. This paper examines the mode effects in these two surveys. The analysis was conducted using cross-tabulations, chi-square tests, and logistic regression models.
|
